![Seerr [Custom Domain Addon]: illustration](https://store.elfhosted.com/wp-content/uploads/2026/10/seerr-custom-domain-addon-card-1-150x150.avif)
OpenClaw
$9.00 / month
Free trial: 7 days
Sign-up fee: $1.00
Hosted OpenClaw, running in its own container on ElfHosted’s infrastructure, away from your laptop and home network. ElfHosted SSO in front, origins pinned to your subdomain, config healed on every start, and the AI included.
A managed instance of OpenClaw (the 365K-star personal AI assistant from openclaw/openclaw) running on ElfHosted in its own isolated container, so the agent’s shell and tools act there and can’t reach your laptop or home LAN. ElfHosted SSO sits in front of the web UI, the Control UI only accepts your own subdomain, and the config is checked and healed on every start. Plus hosting, updates and HTTPS for the channels that need it. MIT-licensed, and the AI is included: the agent works on its first run, with your own provider welcome any time.
OpenClaw is powerful. The Gateway has shell access; channel webhooks deliver untrusted input; tools can browse the web and run code. Running all of that on your own laptop puts your own machine in the blast radius. A hosted instance keeps it in a container on our infrastructure instead.
🔋 AI, batteries included
OpenClaw arrives with an AI model already set up, so the agent works on its very first run. It runs on an ElfHosted model with an allowance included in your plan that covers everyday use, so there is no API key to find. Need more power, a bigger model or heavier use? Bring your own OpenAI, Anthropic, Gemini or other compatible provider in the app’s settings and use it instead, billed by that provider directly.
What is Hosted OpenClaw?
OpenClaw (github.com/openclaw/openclaw) is a 365K-star personal AI assistant. The Gateway is the control plane: a single Node.js daemon that handles channel inbound from WhatsApp / Telegram / Slack / Discord / Signal / iMessage / Microsoft Teams / Matrix / Google Chat / IRC / Mattermost / 13 more, routes messages to the agent, runs tools (browser, shell, sessions, cron), and replies. Voice Wake, Live Canvas, ClawHub skills: all flow through the Gateway. MIT-licensed.
Hosted OpenClaw is that Gateway running on ElfHosted on a public HTTPS subdomain, in its own container. That container is the boundary: the agent’s shell and tools act inside it, not on your devices. ElfHosted SSO sits in front of the web UI, the Gateway’s own token is a per-tenant secret, the Control UI only accepts your subdomain as an origin, and openclaw doctor --fix runs at every start to heal stale config across upgrades. On most DM channels, OpenClaw’s default answers unknown senders with a pairing code before the agent processes anything.
Why OpenClaw Is Hard to Secure on Your Own
OpenClaw’s own security guide doesn’t pretend this is easy. The default-on configuration involves several real choices:
- 🛡️ The Gateway has full host access in your `main` session. That’s intentional: when it’s just you, you want the agent to run shell commands, edit files, send DMs as you. But the same surface is what a compromised agent or a misrouted channel session would inherit.
- 🚪 Channel webhooks are untrusted input. Every Discord DM / Telegram message / Slack mention is text from someone who isn’t you, processed by the agent, with the potential to influence tool calls. Prompt-injection isn’t theoretical here.
- 📦 Sandboxing has to be configured. OpenClaw’s sandboxing is off by default. Its docs show setting `agents.defaults.sandbox.mode` to `”non-main”` with a Docker backend, which means running Docker alongside the Gateway.
- 🔐 The macOS / iOS / Android nodes ask for high-trust permissions. Microphone, accessibility, screen recording, all granted to the OpenClaw app, used by its tooling. Compromise of the host machine is a different threat surface than compromise of a server.
- 🧨 Tools include browser + shell + Discord/Slack actions. The agent can navigate any URL, run any shell command in scope, and send messages as you. A misdirected agent can do real damage.
- 🌐 Public WebSocket exposure is a footgun. Pairing nodes from anywhere in the world is convenient and a risk, so it’s worth opening only the paths you need.
None of these are fatal, and OpenClaw’s docs cover the mitigations. The biggest difference hosting makes is where the agent lives: on ElfHosted it runs in its own container, so whatever it does happens there rather than on the machine holding your files, passwords and home network.
What ElfHosted Sets Up For You
- 🏠 Its own isolated container. Your hosted Gateway runs in its own container on ElfHosted’s infrastructure. The agent’s shell and tools act inside that container, so a misdirected agent doesn’t get a foothold on your laptop or your home network.
- 🔐 ElfHosted SSO in front of the web UI. Defence in depth. The Gateway’s own token, a per-tenant secret, still applies; SSO is a layer on top.
- 📍 Origins pinned to your subdomain. The Control UI only accepts connections from your own ElfHosted subdomain, and forwarded headers are stripped at our edge so the Gateway isn’t fooled about where a request came from.
- 🩺 Config healed on every start.
openclaw doctor --fixruns each time the Gateway starts, so stale settings left behind by an upgrade get repaired instead of piling up. - 🔄 Updates handled. We track upstream OpenClaw releases, so you’re not chasing Node versions or release notes.
- 🔋 Starter AI included. An ElfHosted model is set up on first run, so the agent answers straight away.
- 🔑 Pairing on DM channels. On most DM channels, OpenClaw’s default answers unknown senders with a pairing code, and nothing is processed until you approve them. Keep it that way.
- 🌐 Outside pairing on request. Opening the Gateway WebSocket bridge so devices can pair from outside your network is a hosting-side change our team makes when you ask, for the paths you need.
- ⚠️ One thing to know. OpenClaw’s per-session Docker sandboxing isn’t available on the hosted instance, so group and public-channel sessions share the same container as your own. Treat group channels with care: keep them on pairing or allowlists, and only connect channels and people you trust.
OpenClaw vs Hermes
Hermes Agent is the natural comparison for OpenClaw: both are personal AI agents that can live in chat, use tools, keep useful memory, and run scheduled work without being trapped on your laptop. The difference is emphasis, not winner-takes-all nonsense.
- Choose OpenClaw if you want the OpenClaw Gateway model, broad channel and node support, ClawHub skills, Live Canvas, and a Gateway built around chat channels with pairing on DMs.
- Choose Hermes if you want the Nous Research agent stack, a CLI-first workflow, self-improving skills, strong provider flexibility, and an easy way to experiment with the Hermes ecosystem.
- Try both if you are still figuring out which agent style fits your daily workflow. They solve a similar problem from different angles, and ElfHosted keeps the hosting chores out of the experiment.
OpenClaw is the channel-rich Gateway option. Hermes is the Nous Research, CLI-and-skills-focused option. Either way, the useful part is the same: your agent stays online, reachable, and managed without turning your own machine into the blast radius.
What Hosted OpenClaw Still Does Well
Hosted doesn’t mean unreachable. The agent still answers on the channels you configure, just deliberately, not automatically:
- 💬 One assistant across approved channels. Add Telegram, Slack, or Discord by configuring the channel and approving the senders you want; the agent shows up there with consistent memory and skills.
- 🤖 Slack / Discord bots for your team, with separate sessions per channel. Connect only the channels and people you trust.
- 📞 Voice triggers from the macOS app (and Android Talk Mode): talk to the hosted Gateway, get replies via TTS or back to your channel of choice.
- 🎨 Live Canvas + A2UI for richer agent outputs.
- 📅 Cron jobs & scheduled flows that fire on schedule: the Gateway runs on our servers, so the schedule runs whether you’re at the desk or not.
- 🛠️ Multi-agent routing. Run separate agents for personal DMs vs the team Slack vs an experimental skills sandbox, each with their own workspace.
Technical Specifications
- 🛠️ Software: OpenClaw (FOSS, MIT)
- 🏠 Isolation: the whole Gateway runs in its own container on ElfHosted’s infrastructure; OpenClaw’s per-session Docker sandboxing isn’t available on the hosted instance
- 🔐 DM policy: OpenClaw’s default answers unknown DM senders with a pairing code on most channels
- 🩺 Startup:
openclaw doctor --fixruns at every start to heal stale config - 📡 Channels supported (must be explicitly configured): WhatsApp, Telegram, Slack, Discord, Google Chat, Signal, iMessage (BlueBubbles bridge), Microsoft Teams, Matrix, IRC, Feishu, LINE, Mattermost, Nextcloud Talk, Nostr, Synology Chat, Tlon, Twitch, Zalo, WeChat, QQ, WebChat, plus macOS / iOS / Android nodes
- 🤖 Model: ElfHosted model included; bring your own provider any time, with per-provider failover via OpenClaw’s model config
- 🎙️ Voice: Voice Wake (macOS/iOS), Talk Mode (Android), ElevenLabs TTS + system fallback
- 🧩 Skills: bundled, managed, and workspace skills via ClawHub
- 🔧 Runtime: Node 22.14+ / Node 24, handled by ElfHosted
- 🌐 Access: HTTPS on your own ElfHosted subdomain; ElfHosted SSO in front of the web UI; Control UI origins pinned to your subdomain; outside WebSocket pairing opened on request
- 🔑 Subscription: $9/month, with an ElfHosted model included; your own model subscription is optional
- 🔄 Updates: handled by ElfHosted, tracking upstream OpenClaw point releases
Frequently Asked Questions
Do I need my own API key?
No. Hosted OpenClaw arrives with an ElfHosted model already configured as a provider, so the agent works on its first run. It runs on an allowance included with your plan. Prefer your own model? Add a provider in the usual way and set it as your default: an existing provider or model is never overwritten, and the included one stays in your list alongside it.
How do I run OpenClaw without self-hosting it?
Add Hosted OpenClaw to your ElfHosted subscription. This product is a managed Gateway in its own container, with HTTPS, persistent storage, automatic updates, ElfHosted SSO, origins pinned to your subdomain, and config healed on every start. No `openclaw onboard –install-daemon`, no Node upgrades to babysit.
Is it safer than running OpenClaw on my own machine?
In one important way, yes: the agent runs in its own container on our infrastructure, so even if a prompt-injection attack from a Discord DM convinces it to do something unexpected, it acts inside that container and can’t reach your laptop or your home LAN. Inside the container, though, OpenClaw’s per-session Docker sandboxing isn’t available on the hosted instance, so group and public-channel sessions aren’t separately boxed in. Keep DM channels on pairing, and only connect group channels you trust.
What does ElfHosted set up for me?
Your own isolated container; ElfHosted SSO in front of the web UI; a per-tenant Gateway token; Control UI origins pinned to your subdomain; forwarded headers stripped at our edge; Bonjour discovery off; openclaw doctor --fix at every start; an included starter AI provider; and updates. Opening the WebSocket bridge for outside pairing is something our team does on request. OpenClaw’s own defaults still apply, including pairing codes for unknown DM senders on most channels. OpenClaw’s per-session Docker sandboxing isn’t available on the hosted instance, so we don’t set up sandboxed sessions or tool deny-lists for you.
Why does the Gateway WebSocket need explicit configuration to be public?
Mobile-node pairing across the public internet is useful and also a risk, so we don’t open it unless you ask. Tell us which devices need to pair from outside your network and we’ll configure the WebSocket bridge for those specific paths. Everything else stays closed.
Do I still need OpenAI’s ChatGPT subscription?
No. OpenClaw arrives with an ElfHosted model included, which covers everyday use. If you want a bigger model or heavier use, connect your own ChatGPT Plus, Codex Pro or other provider; that is paid direct to the provider, and the included model stays in your list alongside it.
Does the macOS / iOS app still work with a hosted Gateway?
Yes. Configure the app to point at your ElfHosted subdomain instead of `localhost`. Voice wake, Canvas overlay, push-to-talk all work. Pairing from outside your network needs the WebSocket bridge opened, which our team does on request.
How is this different from running OpenClaw on a Pi or VPS?
Same software, different chores. On a Pi or VPS you handle the cert renewal, the Node upgrades, the updates and the config clean-up after each one. On ElfHosted, those are handled, and openclaw doctor --fix runs at every start so upgrades don’t leave stale config behind. One trade-off: a Pi or VPS can run Docker for OpenClaw’s per-session sandboxing, while that sandboxing isn’t available on the hosted instance.
Can I run multiple agents (personal vs team)?
Yes. OpenClaw’s multi-agent routing isolates different channels/peers to separate workspaces. Common setup: personal DMs in one agent, team Slack in another, experimental skills in a third. They share your container, so connect only channels you trust.
Where do I get help?
The official OpenClaw Discord for upstream questions; the ElfHosted Discord for hosting and security questions specific to this managed instance.
Hosted OpenClaw is a managed Gateway in its own isolated container, away from your devices, with ElfHosted SSO in front, origins pinned to your subdomain, config healed on every start, and updates handled. Your agent keeps its shell access and inbox webhooks, and they act on our infrastructure instead of your laptop. $9/month for the hosting, with the AI included. EXFOLIATE!
Only logged in customers who have purchased this product may leave a review.
Reviews
Clear filtersThere are no reviews yet.